The influence of grassland restoration and weather patterns on avian species within the coastal prairie ecoregion

Grassland obligate birds are experiencing population declines across North America. Habitat loss due to brush encroachment is a primary driver of these declines. Considerable attention has been focused on prairie restoration and avian conservation in the Great Plains region of North America; far less attention has been given to the coastal prairies paralleling the Gulf of Mexico. This thesis reflects the beginning steps of a restoration project in which the goal is to restore a brush encroached site to a native coastal prairie community on the Welder Wildlife Foundation (WWF) in Sinton, Texas. Chapter I contains a literature review on various restoration tools aimed at grassland restoration and avian conservation. In chapters II and III, I cover the initiation of the WWF restoration project where I assessed the vegetative and avian community response to a brush-targeted herbicide treatment aimed at reducing mesquite (Prosopis glandulosa) and huisache (Vachellia farnesiana) cover. Interesting observations of the dickcissel (Spiza americana) on my study plots warranted further exploration of the species distributional patterns. In chapter IV, I used Breeding Bird Survey data to assess the relationship of weather variables on dickcissel abundance within the coastal prairie ecoregion from 1980-2016. I conclude this thesis in chapter V with final thoughts and directions for the WWF restoration project. For chapter II and III, I used a before/after–control/impact study design and established 58 survey points in a 144-ha treated plot and a 216-ha control plot. From these points, I conducted avian point counts and vegetation surveys during the summer (May-June) prior to the Aminocyclopyrachlor herbicide application in October 2014, and continued post-treatment surveys for the breeding seasons (May-July) and wintering seasons (January) of 2015, 2016, and 2017. In chapter II, I assessed changes in brush and herbaceous percent-cover within both plots during the study. Herbicide treatment was effective in significantly reducing mesquite and huisache cover, with no regrowth among huisache and single-stemmed mesquite trees. Graminoid cover significantly increased in both plots during the study, with a significant increase among invasive grasses in the treated plot. Above average rainfall in the spring following herbicide application was a confounding factor that had parallel effects on the herbaceous community, particularly grasses, across the study site. I suspect this also impacted the effectiveness of herbicide on low-growth, multi-stemmed mesquite trees, as there was observable regrowth on in subsequent years. Woody structures remained on the treated plot following treatment, and additional brush removal treatments are required to continue with restoration of the site. In chapter III, I assessed changes in avian species diversity, community similarity, guild structure, and species density in response to vegetative changes covered in chapter II. Breeding bird species diversity and woody-associated species density decreased in the treated plot following herbicide application, with a notable increase in grassland species representation due to the novel occurrence of dickcissels breeding on both study sites. Wintering bird species diversity was variable year-to-year yet displayed a net decrease during the study period. Additionally, woody-associated species density decreased during the study. Changes in breeding and wintering bird community similarity were lower in magnitude from year to year, but net changes from the beginning and ending of the study period were evident. These changes were evident among woody-associated species such as the northern cardinal (Cardinalis cardinalis), white-eyed vireo (Vireo griseus), and painted bunting (Passerina ciris). The arrival of dickcissels in large numbers on the study site following above average rainfall prompted further exploration of the species relationship with weather, particularly precipitation. In chapter IV, I used an analysis of covariance to determine if dickcissel abundance correlates with different levels of rainfall and temperature. My results indicated that below normal, dry season rainfall had a negative impact on dickcissel abundance, potentially due to the reduction of vegetative cover and arthropod abundance. The mechanism behind the dickcissel occurrence is unclear, however condition on the coastal prairie might lead to opportunistic breeding in years with above normal, dry-season rainfall. However, more study is needed. In conclusion, this thesis reflects the importance of time, environmental conditions, and flexibility when assessing changes at the beginning of a restoration project on the WWF. Changes in bird communities may not be evident in the year following restoration efforts, but changes across larger periods of time might reveal patterns related to restoration. This is relevant during years when variable environmental conditions mask patterns of change related to restoration efforts. Ultimately, adaptive management aimed at addressing variable treatment outcomes such as mesquite resprout is necessary to overcome community resistance and cross the critical threshold between the brush encroached and desired grassland state.